Theory

• Pooling

Layer 1

Layer 2/3

High order sequence memory

Inputs

• Output from Layer 4
• Output from Layer 1 (column space)

Output

• Layer 4 of another region
• Layer 5 (column activations)

Layer 4

Sensorimotor transitions

Inputs

• Union of sensory and motor data

Outputs

• If prediction is correct pass prediction to Layer 3
• If prediction is incorrect pass through signal from the input

Layer 5

Recall motor sequences. Generates behavior

Inputs

• Layer 3. Copy of motor commands.
• Output from Layer 1 (column space)

Outputs

• Motors, Switches, Controllers. The axons of the Layer 5 try to connect with dendrites of cell's in the old brain, who's axons generated behavior and send information to Layer 4.
• Layer 4 of higher region

Layer 6

Attention

Input

• Union of sensory and motor data

Output

• to Layer 5

Problems

1. Dimensions of Layer 4 input and output are different. How do we pass the information through?